Ribose operon repressor (RbsR) contributes to the adhesion of Aeromonas hydrophila to Anguilla japonica mucus

Abstract The characterization of adhesion between pathogenic bacteria and the host is critical. Pathogenic Aeromonas hydrophila was shown to adhere in vitro to the mucus of Anguilla japonica. To further investigate the adhesion mechanisms of A. hydrophila, a mini‐Tn10 transposon mutagenesis system was used to generate an insertion mutant library by cell conjugation. Seven mutants that were impaired in adhesion to mucus were selected out of 332 individual colonies, and mutant M196 was the most severely impaired strain. National Center for Biotechnology Information (NCBI) blast analysis showed that mutant M196 was inserted by mini‐Tn10 with an ORF of approximately 1,005 bp (GenBank accession numbers KP280172). This ORF is predicted to encode a protein consist of 334 amino acid, which displays the highest identity (98%) with the RbsR of A. hydrophila ATCC 7966. Random inactivation of rbsR gene affected the pleiotropic phenotypes of A. hydrophila. The adhesion ability and the survival level of the rbsR gene mutant (M196) were attenuated compared with the wild‐type and rbsR complementary type. The findings of this study indicated that RbsR plays roles in the bacterial adhesion and intracellular survival of A. hydrophila.


| INTRODUCTION
Adhesion to the host external surfaces and tissues by pathogenic bacteria represents the first crucial step in most infections, and it had been regarded as an important virulence factor in colonizing the host surface for proliferation and for the release of virulence factors of many pathogenic bacteria (Grześkowiak, Collado, Vesterlund, Mazurkiewicz, & Salminen, 2011;Huang et al., 2016;Wang et al., 2015).The ability of pathogens to attach to the host mucus, cells, and other tissues is mediated by adhesions that can cause many infectious diseases or other harm (Hori & Matsumoto, 2010). Therefore, the inhibition of bacterial adhesion to the host could be useful therapeutically (Acord, Maskell, & Sefton, 2005). It is important to understand the mechanisms of bacterial adhesion as well as the major factors controlling this attachment.
Recently, the adhesion of many clinical pathogens to their hosts has been elucidated, and bacterial adhesion was shown to be mostly influenced by motility, cell hydrophobicity, and bacterial surface structures (Chen, Yan, Wang, Zhuang, & Wang, 2008;He et al., 2015;Losensky, Vidakovic, Klingl, Pfeifer, & Fröls, 2015). Experiments with the pathogens Pseudomonas aeruginosa, Vibrio alginolyticus; Vibrio vulnificus, and Legionella pneumophila (Huang, Hu et al., 2015;Kim & Rhee, 2003;Luo et al., 2016;Olga, Jens, & Michael, 2011;Shao et al., 2013) demonstrated the important role of flagella or lipopoly saccharides in establishing the initial interaction with mucosal surfaces or cells and that this defect can reduce the ability of adhesion. In addition, the genetic basis of colonization has been studied in V. vulnificus, L. pneumophila, and Streptococcus suis (Chang et al., 2005;Kim & Rhee, 2003;Zhang et al., 2015) and in the genes that impact bacterial adhesion that can also contribute to bacterial pathogenesis. In the pathogenic bacteria V. vulnificus, FlgC was shown to be involved in adhesion to HeLa cells, and the flgC mutation could decrease the level of adhesion, cytotoxicity, and lethality in mice (Kim & Rhee, 2003). Intracellular bacteria L. pneumophilalaia was shown to be involved in adhesion to the human lung alveolar epithelial cell line A549, and the laiA gene mutation could cause reduced mortality in A/J mice (Chang et al., 2005). The dnaJ of S. suis Type 2 was important for adhesion to host cells, and the dna-JPKO strain can result in reduced adhesion to cultured cell monolayers (Zhang et al., 2015). These pathogenic genes were all considered to be the key factors for bacterial adhesion and pathogenicity.
Aeromonas hydrophila has been recognized as a gram-negative bacillus of the Aeromonadaceae family, and it can survive in a wide variety of aquatic systems (Cao et al., 2012). It is an opportunistic pathogen and can infect a great variety of poikilothermic and homeothermic animals, including humans (Krovacek et al., 1994;Mccoy et al., 2010;Sime-Ngando, 2015). Several structures have previously been demonstrated to be related to the pathogenicity of A. hydrophila, including adhesion (Grześkowiak et al., 2011;Hori & Matsumoto, 2010;Van Der Marel, Schroers, Neuhaus, & Steinhagen, 2008). A. hydrophila attached strongly to the host surface using flagella-promoted motility in the optimum environment, and the environmental factors obviously affected bacterial adhesion (Benhamed, Guardiola, Mars, & Esteban, 2014). Several genes essential for A. hydrophila adhesion have been identified. The A. hydrophila flgE gene mutation can cause the absence of flagella, and the mutant bacteria can exhibit inadequate motility, adhesion, invasion, and survival in host macrophages when compared with the wild-type B11 and the complementary strain (Qin et al., 2014). Single gene mutations of minD can lead to flagella deficiency and adhesion reduction .
It has been extensively reported that some genes associated with flagella play an important role in A. hydrophila adhesion. However, few studies have focused on fully exploring the adhesion-related genes of the pathogenic bacteria A. hydrophila, especially genes related to nutrient metabolism or enzymes. The aim of this study was to identify more adhesion-related genes of pathogenic A. hydrophila with random gene inactivation method and to probe other pleiotropic phenotypes of A. hydrophila by constructing mutant and complementary strains to study the mechanisms of these genes.

| Bacterial strains and growth conditions
The bacterial strains and plasmids used in this study are listed in

| Preparation of A. japonica mucus and A. hydrophila adhesion in mucus in vitro
A group of five healthy A. japonica were purchased from an aquatic product market in Xiamen of Fujian Province. The skin mucus was prepared using a modification of a previously described technique . Briefly, the skin mucus gel was collected by scraping the skin surfaces with a rubber spatula, and the gel was homogenized in PBS and centrifuged twice at 20,000 g at 4°C for 30 min to remove particulates. The supernatant was filtered sequentially through .45 and .2 μm pore size filters. The resulting mucus sample was adjusted to 1 mg of protein/ml with sterile aged seawater, and the protein concentration was determined using the method that was described by Bradford (Bradford, 1976).
Gill mucus was prepared using a technology we described before (Chen et al., 2008). Blood was completely removed from the caudal vessel, and the gill arches were excised and soaked in PBS for 2 hr at 4°C, with occasional shaking. The mucus preparation was centrifuged twice at 20,000 g for 30 min at 4°C to remove particles and cellular material, followed by filtration of the final supernatant through .45 and .2 μm pore size filters. The protein concentration of mucus preparation was adjusted to 1 mg/ml in PBS. The protein concentration was determined using the method of Bradford (1976).
Intestinal mucus was prepared using a technique we described before (Yan, Chen, Ma, Zhuang, & Wang, 2007). The fish were starved for 48 hr without feeding and the intestines were then removed and divided into foregut and hindgut according to intestinal morphology.
The guts were transferred to sterile petri dishes and washed by sterile .01 mol/L PBS (pH 7.2). Next, they were split open with a scalpel, and the foregut and hindgut mucus were collected by scrapping of the inner surface of the intestines with a rubber spatula to remove the layer of mucus gel covering the intestinal lumen, respectively.
The mucus was homogenized in .01 mol/L PBS (pH 7.2).The mucous preparations were centrifuged twice at 20,000 g for 30 min at 4°C to remove particulate and cellular material, followed by filtration of the final supernatant through .45-and .2μm pore size filters. The mucus samples were adjusted to 1 mg protein/ml with PBS. The protein concentration was determined using the method of Bradford (1976).
An in vitro adhesion assay was performed using an indirect enzyme-linked immunosorbent assay (ELISA) method with 96-well microtiter polystyrene plates, as previously described and modified (Qin et al., 2014). Briefly, 100 μl of the mucus samples were added to each well, the mucus was fixed by incubation overnight at 4°C, and

| Mini-Tn10 mutagenesis of A. hydrophila strain W1 and genetic screening for adhesion-deficient mutants
The mini-

| Southern blotting
A single insertion event in the rbsR::Km mutant was confirmed. A single chromosomal transposon insertion in the rbsR::Km mutant was confirmed with genomic DNA prepared from the rbsR::Km mutant using the standard method (Rock & Nelson, 2006

| Sequence analysis of the mini-Tn10 insertion mutants
Thermal asymmetric interlaced polymerase chain reaction (TAIL-PCR) (Singer & Burke, 2003) was used to obtain the mutant DNA sequences flanking the transposon to find the insertion sites in A. hydrophila. The arbitrary primers were provided by a genomic walking kit (Takara, Bio).
The specific transposon nested primers are listed in Table 2    to incubate at 28°C in 5% CO 2 with the time point denoted as 0 hr.

| Preparation of macrophages, and in vitro invasion and survival
After incubation for 1 hr, the cells were centrifuged for 5 min at 400 g at 28°C, and the supernatant was aspirated. Then 1 ml of sterile distilled water was added for 30 min to lyse the cells. The CFU number of the cell lysate was determined by plate counting.

| Statistical analyses
All data were statistically analyzed with SPSS16.0 (SPSS, Chicago, IL, USA). The mean ± standard deviation was calculated for each sample. Unless otherwise stated, all experiments were performed at least three times in triplicate assays.

| Adhesion-reduced mutant isolation
A mini-Tn10 transposon mutagenesis system was used to generate an insertion mutant library by cell conjugation. A mutagenesis library with 332 random insertion mutants of A. hydrophila was constructed, and each mutant in the library was subjected to the adhesion assay. Seven mutants were found to be reduced in adhesion to A. japonica skin mucus compared with W1, and the mutant strain (M196) was the most significantly inadequate regarding adhesion to A. japonica mucus (Figure 1).

| Southern blotting
Southern blotting analysis showed that a single band was present in all mutants and the plasmid positive control pLOF/Km. No signal was detected in the negative control wild-type strain W1 (Figure 2), which confirmed that the mutations were caused by the insertion of transposon mini-Tn10Km and that only a single transposon insertion was present in the chromosome for each mutant.

| Mini-Tn10 insertion site in the mutant strain (M196)
Sequence analysis showed that the mini-Tn10 insertion site was a 1,005 bp ORF that shared the highest identity (98%) with the rbsR gene of A. hydrophila ATCC 7966 (accession no: CP000462.1). This rbsR gene encoded a ribose operon, which is a repressor protein that was predicted to contain 334 amino acid residues with a predicted 33 kDa molecular weight. The Tn10 insertion site is indicated with a bold and enlarged letter (Figure 3).

| RbsR protein expression in the complementary strain
After the recombinant expression plasmid pACYC184-rbsR was introduced into the mutant (M196), the expression of protein RbsR in the complementary strain (MC196) was detected by western blotting.
Western blotting analysis showed that a positive signal of approximately 33 kDa was detected, which confirmed that the fusion protein RbsR-HA was expressed in the complementary strain (MC196) but not in the mutant strain (M196; Figure 4).

| Adhesion assay
Adhesion of the wild-type (W1), the rbsR mutant strain (M196), and the complementary strain (MC196) to the host skin, intestinal and gill mucus was detected. The results showed that the adhesion of the M196 bacteria to skin, intestinal and gill mucus was approximately 15%, 46% and 14%, respectively, compared with that of the wild-type (W1) strain. The adhesion of bacteria to the skin, intestinal and gill mucus with the complementary strain recovered to more than 43%, 59%, and 35% of that of the wild-type strain, respectively ( Figure 5).
These data suggested that the mutation in rbsR significantly affected bacterial adhesion.

| Intracellular survival in macrophages
Given were ~26% and ~31%, respectively, which suggested that the survival of the mutant bacteria in host macrophages were seriously impaired.

| DISCUSSION
Recently, adhesion has been found to be associated with the virulence of many types of pathogens (Benhamed et al., 2014;Luo et al., 2016). Some studies have suggested that bacterial adhesion genes, such as flrA, flrB, and flrC can mostly be related to flagella. Bacterial adhesion is an extremely complicated process and can be controlled by one or more genes. To understand the genetic basis of adhesion in the A. hydrophila prototype, transposon insertion mutants that were inadequate in adhesion isolated were characterized. In this study, pathogenic A. hydrophila W1 could adhere to A. japonica mucus well, and a gene corresponding to a repressor protein was identified: rbsR Previous studies have found a few genes, such as CtBP, which is not traditionally considered as an adhesion factor, that can impact adhesion (Grooteclaes & Frisch, 2000). The results confirmed the screening data in this study.
In this study, the mutant strain (M196) was significantly inadequate regarding adhesion to A. japonica mucus and the complementary strain (MC196) recovered the adhesion ability to some extent. rbsR deletion can immediately affect the mRNA level of the whole rbsRACBD (rbs) operon, and the gene products of the rbs operon of Corynebacterium glutamicum encode a ribose-specific ATP-binding cassette (ABC) transport system with corresponding regulatory proteins (i.e., rbsR).
ABC transporters are membrane proteins that couple the energy from ATP hydrolysis to transport substrates, including nutrients and toxins (Chen, Lu, Lin, Davidson, & Quiocho, 2003;Nentwich et al., 2009 protein has been believed to be an ordinary LacI-type repressor for a single target rbsDACBK operon, which may be involved in the production of GMP in nucleotide synthesis (Shimada, Kori, & Ishihama, 2013). Mostly, the E. coli rbsR gene was found to be a transcriptional repressor for the ribose operon that is involved in the regulation of purine nucleotide synthesis and nutrient synthesis and transport (Daigle, Graham, & Curtiss, 2001;Shimada et al., 2013). In this study, the data suggested that the survival ability of the rbsR mutant in host macrophages was severely impaired. This impairment could be related to the result that Salmonella typhi rbsR mutations could decrease the level of survival within human macrophages at both 2 and 24 hr after bacterial infection (Ocallaghan, Maskell, Liew, Easmon, & Dougan, 1988